Back to main page

Keeping fingerprints in sync with your compound data

After installing OrChem and creating the initial fingerprints, these fingerprints will get out of date over time if your compound data changes after inserts, deletes and updates. OrChem provides a simple mechanism that can be used to periodically synchronize the compounds with the fingerprint, the details are described on this page.

Set up

The synchronization concept is pretty straightforward: a CRUD trigger is created on your compound base table and each change is then logged in an audit table specifically for OrChem. Periodically (at the interval you want), you then schedule an OrChem procedure to inspect the audit table, update the fingerprints based on the audit table content and remove processed audit data.
This differs a little bit from the general installation that was done in its own "ORCHEM" schema. The trigger and audit table must be installed in the 'master' schema that owns the compound table.

  • Navigate into the orchem/sql directory where you unzipped OrChem
  • Connect to the database using SQL*Plus with your account that owns the compound table.
  • In SQL*Plus, execute the script synchronization.sql, located in orchem/sql. This script first creates the audit table and then asks you for the compound table name and the name of the primary key column. It then creates a trigger on your compound table.
    Example below, where a trigger is created on compound table 'ORCHEM_COMPOUND_SAMPLE' which has a primary key column called 'ID'.
            SQL> @synchronization
            drop table orchem_audit_compound_changes (if exists)
            Table dropped.
            creating table orchem_audit_compound_changes
            Table created.
            creating trigger 'orchem_trg_compound_changes'
            Enter value for your_compound_table: ORCHEM_COMPOUND_SAMPLE
            Enter value for your_primary_key: ID
            Trigger created.
            No errors.
  • With the creation of the trigger, changes on the compound table will be logged in a new audit table called "orchem_audit_compound_changes". The OrChem user needs to be given privileges in order to use this audit table. The privilege is given with a GRANT command.

    Like the previous script, the GRANT needs to run by the account that owns the compound table (and the newly created audit table). You can copy/paste/execute the example below, assuming your OrChem account is actually called "ORCHEM"; otherwise you have to change the last word, the grantee.
            SQL> grant select,insert, update, delete on orchem_audit_compound_changes to orchem;
            Grant succeeded.
  • Now reconnect to the database using SQL*Plus with your account ORCHEM account instead.
  • To finalize the setup, OrChem needs to create a private synonym for the audit table. This makes it possible for Orchem to refer to the audit table without a schema prefix.
    You can use the example below, replacing xxx with the name of the schema in which you just created the audit table. If you don't have the privilege to create synonyms, you'll get an error (so go and bug your DBA)
            SQL>  create synonym orchem_audit_compound_changes for xxxx.orchem_audit_compound_changes;
            Synonym created.
  • This complete the synchronization set up.

Running the synchronization

Synchronization is invoked by calling package "orchem_fingerprinting", procedure "synchronize_cdk_fingerprints". An example is given below, in this case the procedure returns immediately because there was nothing to do (the audit table was empty).

SQL> exec orchem_fingerprinting.SYNCHRONIZE_CDK_FINGERPRINTS

Nothing to do !!
PL/SQL procedure successfully completed.

Elapsed: 00:00:00.12

You may want to schedule this procedure regularly, depending on the volatility of you compound data. For such scheduling you could use the standard facility DBMS_JOB. Or you could run it each time after you update/load compounds in some batch job.

Back to main page